
There are few complaints from my side - I've had a stable system with 2 Pulsar Ones, then I bought that 'supposed to be broken' 3rd Pulsar for a bargain.
Of course I was happy that I could just add it and had some additional DSPs.
That system never ran out of sync, it's on a 1.2 Tualatin Asus TUSL2.
For testing purposes I sometimes remove the cards and plug them into something else - in this case an old HP workstation with dual P3.
I didn't dare to add the third card as it would end only 2 millimeters from the heatsinks of the HP 3D card, which is passively 'cooled' and hot as hell...
That's how I found out that under certain conditions (this is a dual CPU server board) the 'last' Pulsar has some trouble - possibly the previous owner had a dualcore machine, too - I don't know if the card reacts the same in the Asus board, it's 'regular' home, it might be more tolerant...
The area around some capacitors on the circuit board does look suspective indeed, so I thought I'd replace them and see if it improves things.
I ran out of parts yesterday and only replaced 4, it seems to have an infkuence as it seems to take longer for the dropout to appear and quicker to recover.
One should always measure and write such findings down... stupid me.
But the board is a bitch to (de)solder - it's extremely sensitive

... and yes, now I should know about pain, but there's this Hoyer 335 copy with the slightly high action and it's steel strings torturing my fingertips...

[later]
replacing the caps didn't change anything
moving the Pulsars back to the Asus all sync is stable and rocksolid.
nevertheless it was quite interesting, as on the dual cpu HP (under win2k server) SFP 'forgot' about the registration a couple of times when removing and adding a board.
Pointed it to the keyfile and the imported succeeded - rebooting etc didn't change a thing, but when one of the boards was exchanged the request was up again - strange, as the Pulsar related to the keys was (of course) never removed

I don't have a clue because the respective system was tested once with 2 Pulsars and worked, but since it's PCI performance was rather disappointing I SFP got it's dedicated Asus TUSL (leaving the SFP installation on the HP)
[/later]
cheers, Tom